Loan_prediction.ipynb

{
  "cells": [
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "import os\nimport pandas as pd\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport warnings                        # To ignore any warnings\nwarnings.filterwarnings(\"ignore\")",
      "execution_count": 1,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Data EDA and Importing Data"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train=pd.read_csv('train_u6lujuX_CVtuZ9i.csv')\ntest=pd.read_csv('test_Y3wMUE5_7gLdaTN.csv')\n\ntrain_original=train.copy() \ntest_original=test.copy()\n\ntrain.info()",
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 614 entries, 0 to 613\nData columns (total 13 columns):\nLoan_ID              614 non-null object\nGender               601 non-null object\nMarried              611 non-null object\nDependents           599 non-null object\nEducation            614 non-null object\nSelf_Employed        582 non-null object\nApplicantIncome      614 non-null int64\nCoapplicantIncome    614 non-null float64\nLoanAmount           592 non-null float64\nLoan_Amount_Term     600 non-null float64\nCredit_History       564 non-null float64\nProperty_Area        614 non-null object\nLoan_Status          614 non-null object\ndtypes: float64(4), int64(1), object(8)\nmemory usage: 62.4+ KB\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.isnull().sum()",
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 3,
          "data": {
            "text/plain": "Loan_ID               0\nGender               13\nMarried               3\nDependents           15\nEducation             0\nSelf_Employed        32\nApplicantIncome       0\nCoapplicantIncome     0\nLoanAmount           22\nLoan_Amount_Term     14\nCredit_History       50\nProperty_Area         0\nLoan_Status           0\ndtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test.isnull().sum()",
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 4,
          "data": {
            "text/plain": "Loan_ID               0\nGender               11\nMarried               0\nDependents           10\nEducation             0\nSelf_Employed        23\nApplicantIncome       0\nCoapplicantIncome     0\nLoanAmount            5\nLoan_Amount_Term      6\nCredit_History       29\nProperty_Area         0\ndtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.describe()",
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 5,
          "data": {
            "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>ApplicantIncome</th>\n      <th>CoapplicantIncome</th>\n      <th>LoanAmount</th>\n      <th>Loan_Amount_Term</th>\n      <th>Credit_History</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>count</th>\n      <td>614.000000</td>\n      <td>614.000000</td>\n      <td>592.000000</td>\n      <td>600.00000</td>\n      <td>564.000000</td>\n    </tr>\n    <tr>\n      <th>mean</th>\n      <td>5403.459283</td>\n      <td>1621.245798</td>\n      <td>146.412162</td>\n      <td>342.00000</td>\n      <td>0.842199</td>\n    </tr>\n    <tr>\n      <th>std</th>\n      <td>6109.041673</td>\n      <td>2926.248369</td>\n      <td>85.587325</td>\n      <td>65.12041</td>\n      <td>0.364878</td>\n    </tr>\n    <tr>\n      <th>min</th>\n      <td>150.000000</td>\n      <td>0.000000</td>\n      <td>9.000000</td>\n      <td>12.00000</td>\n      <td>0.000000</td>\n    </tr>\n    <tr>\n      <th>25%</th>\n      <td>2877.500000</td>\n      <td>0.000000</td>\n      <td>100.000000</td>\n      <td>360.00000</td>\n      <td>1.000000</td>\n    </tr>\n    <tr>\n      <th>50%</th>\n      <td>3812.500000</td>\n      <td>1188.500000</td>\n      <td>128.000000</td>\n      <td>360.00000</td>\n      <td>1.000000</td>\n    </tr>\n    <tr>\n      <th>75%</th>\n      <td>5795.000000</td>\n      <td>2297.250000</td>\n      <td>168.000000</td>\n      <td>360.00000</td>\n      <td>1.000000</td>\n    </tr>\n    <tr>\n      <th>max</th>\n      <td>81000.000000</td>\n      <td>41667.000000</td>\n      <td>700.000000</td>\n      <td>480.00000</td>\n      <td>1.000000</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
            "text/plain": "       ApplicantIncome  CoapplicantIncome  LoanAmount  Loan_Amount_Term  \\\ncount       614.000000         614.000000  592.000000         600.00000   \nmean       5403.459283        1621.245798  146.412162         342.00000   \nstd        6109.041673        2926.248369   85.587325          65.12041   \nmin         150.000000           0.000000    9.000000          12.00000   \n25%        2877.500000           0.000000  100.000000         360.00000   \n50%        3812.500000        1188.500000  128.000000         360.00000   \n75%        5795.000000        2297.250000  168.000000         360.00000   \nmax       81000.000000       41667.000000  700.000000         480.00000   \n\n       Credit_History  \ncount      564.000000  \nmean         0.842199  \nstd          0.364878  \nmin          0.000000  \n25%          1.000000  \n50%          1.000000  \n75%          1.000000  \nmax          1.000000  "
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "category_columns=train.select_dtypes(include=['object']).columns\ncategory_columns",
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 6,
          "data": {
            "text/plain": "Index(['Loan_ID', 'Gender', 'Married', 'Dependents', 'Education',\n       'Self_Employed', 'Property_Area', 'Loan_Status'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "nums=train.select_dtypes(include=['int64','float64']).columns\nnums",
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 7,
          "data": {
            "text/plain": "Index(['ApplicantIncome', 'CoapplicantIncome', 'LoanAmount',\n       'Loan_Amount_Term', 'Credit_History'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.columns",
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 8,
          "data": {
            "text/plain": "Index(['Loan_ID', 'Gender', 'Married', 'Dependents', 'Education',\n       'Self_Employed', 'ApplicantIncome', 'CoapplicantIncome', 'LoanAmount',\n       'Loan_Amount_Term', 'Credit_History', 'Property_Area', 'Loan_Status'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "len(train.columns)",
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 9,
          "data": {
            "text/plain": "13"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "len(nums)+len(category_columns)",
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 10,
          "data": {
            "text/plain": "13"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Let us check out type of values in category"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "category_columns",
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 11,
          "data": {
            "text/plain": "Index(['Loan_ID', 'Gender', 'Married', 'Dependents', 'Education',\n       'Self_Employed', 'Property_Area', 'Loan_Status'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.shape",
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 12,
          "data": {
            "text/plain": "(614, 13)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "len(train.Loan_ID.value_counts())",
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 13,
          "data": {
            "text/plain": "614"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Gender.value_counts()",
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 14,
          "data": {
            "text/plain": "Male      489\nFemale    112\nName: Gender, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Dependents.value_counts()",
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 15,
          "data": {
            "text/plain": "0     345\n1     102\n2     101\n3+     51\nName: Dependents, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Education.value_counts()",
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 16,
          "data": {
            "text/plain": "Graduate        480\nNot Graduate    134\nName: Education, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Self_Employed.value_counts()",
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 17,
          "data": {
            "text/plain": "No     500\nYes     82\nName: Self_Employed, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Property_Area.value_counts()",
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 18,
          "data": {
            "text/plain": "Semiurban    233\nUrban        202\nRural        179\nName: Property_Area, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Loan_Status.value_counts()",
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 19,
          "data": {
            "text/plain": "Y    422\nN    192\nName: Loan_Status, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['Dependents'].replace('3+', 3,inplace=True) \ntest['Dependents'].replace('3+', 3,inplace=True) \ntrain['Loan_Status'].replace('N', 0,inplace=True) \ntrain['Loan_Status'].replace('Y', 1,inplace=True)",
      "execution_count": 20,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Loan_Status.value_counts()",
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 21,
          "data": {
            "text/plain": "1    422\n0    192\nName: Loan_Status, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.Dependents.value_counts()",
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 22,
          "data": {
            "text/plain": "0    345\n1    102\n2    101\n3     51\nName: Dependents, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test.Dependents.value_counts()",
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 24,
          "data": {
            "text/plain": "0    200\n2     59\n1     58\n3     40\nName: Dependents, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train[category_columns].isnull().sum()",
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 23,
          "data": {
            "text/plain": "Loan_ID           0\nGender           13\nMarried           3\nDependents       15\nEducation         0\nSelf_Employed    32\nProperty_Area     0\nLoan_Status       0\ndtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "We can consider these methods to fill the missing values:\n\n**For numerical variables: imputation using mean or median**\n\n**For categorical variables: imputation using mode(occuring max times)**"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['Gender'].fillna(train['Gender'].mode()[0], inplace=True) \ntrain['Married'].fillna(train['Married'].mode()[0], inplace=True)\ntrain['Dependents'].fillna(train['Dependents'].mode()[0], inplace=True)\ntrain['Self_Employed'].fillna(train['Self_Employed'].mode()[0], inplace=True)\ntrain['Credit_History'].fillna(train['Credit_History'].mode()[0], inplace=True)",
      "execution_count": 25,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train[category_columns].isnull().sum()",
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 26,
          "data": {
            "text/plain": "Loan_ID          0\nGender           0\nMarried          0\nDependents       0\nEducation        0\nSelf_Employed    0\nProperty_Area    0\nLoan_Status      0\ndtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test['Gender'].fillna(train['Gender'].mode()[0], inplace=True) \ntest['Dependents'].fillna(train['Dependents'].mode()[0], inplace=True)\ntest['Self_Employed'].fillna(train['Self_Employed'].mode()[0], inplace=True)\ntest['Credit_History'].fillna(train['Credit_History'].mode()[0], inplace=True)\ntest['Loan_Amount_Term'].fillna(train['Loan_Amount_Term'].mode()[0], inplace=True)\ntest['LoanAmount'].fillna(train['LoanAmount'].median(), inplace=True)\n",
      "execution_count": 27,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test.isnull().sum()",
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 28,
          "data": {
            "text/plain": "Loan_ID              0\nGender               0\nMarried              0\nDependents           0\nEducation            0\nSelf_Employed        0\nApplicantIncome      0\nCoapplicantIncome    0\nLoanAmount           0\nLoan_Amount_Term     0\nCredit_History       0\nProperty_Area        0\ndtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## We have treated the categorical null values now we will treat the numerical values"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train[nums].isnull().sum()",
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 29,
          "data": {
            "text/plain": "ApplicantIncome       0\nCoapplicantIncome     0\nLoanAmount           22\nLoan_Amount_Term     14\nCredit_History        0\ndtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test[nums].isnull().sum()",
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 31,
          "data": {
            "text/plain": "ApplicantIncome      0\nCoapplicantIncome    0\nLoanAmount           0\nLoan_Amount_Term     0\nCredit_History       0\ndtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['Loan_Amount_Term'].value_counts()\n",
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 30,
          "data": {
            "text/plain": "360.0    512\n180.0     44\n480.0     15\n300.0     13\n84.0       4\n240.0      4\n120.0      3\n36.0       2\n60.0       2\n12.0       1\nName: Loan_Amount_Term, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "#It can be seen that in loan amount term variable, the value of 360 is repeating the most. \n#So we will replace the missing values in this variable using the mode of this variable.\ntrain['Loan_Amount_Term'].fillna(train['Loan_Amount_Term'].mode()[0], inplace=True)",
      "execution_count": 32,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['LoanAmount'].hist(bins=20) ",
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 33,
          "data": {
            "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x7fd0ea4f9898>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### LoanAmount has outliers"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "#As Loan amount has lot of outliers so mean is not a good approach to fill missing values\ntrain['LoanAmount'].fillna(train['LoanAmount'].median(), inplace=True)",
      "execution_count": 34,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Outlier Treatment for both test and train data"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['LoanAmount'].hist(bins=20) ",
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 35,
          "data": {
            "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x7fd0ea1c04a8>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test['LoanAmount'].hist(bins=20) ",
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 36,
          "data": {
            "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x7fd0ea172748>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['LoanAmount_log'] = np.log(train['LoanAmount'])",
      "execution_count": 37,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['LoanAmount_log'].hist(bins=20) ",
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 38,
          "data": {
            "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x7fd0ea1b6e80>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test['LoanAmount_log'] = np.log(test['LoanAmount'])",
      "execution_count": 39,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['LoanAmount_log'].hist(bins=20) ",
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 40,
          "data": {
            "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x7fd0ea054da0>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Dropping uncessary columns and splitting data"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.columns",
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 41,
          "data": {
            "text/plain": "Index(['Loan_ID', 'Gender', 'Married', 'Dependents', 'Education',\n       'Self_Employed', 'ApplicantIncome', 'CoapplicantIncome', 'LoanAmount',\n       'Loan_Amount_Term', 'Credit_History', 'Property_Area', 'Loan_Status',\n       'LoanAmount_log'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train=train.drop(['Loan_ID','LoanAmount'],axis=1) ",
      "execution_count": 42,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.columns",
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 43,
          "data": {
            "text/plain": "Index(['Gender', 'Married', 'Dependents', 'Education', 'Self_Employed',\n       'ApplicantIncome', 'CoapplicantIncome', 'Loan_Amount_Term',\n       'Credit_History', 'Property_Area', 'Loan_Status', 'LoanAmount_log'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.shape",
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 44,
          "data": {
            "text/plain": "(614, 12)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test=test.drop(['Loan_ID','LoanAmount'],axis=1)",
      "execution_count": 45,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test.columns",
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 46,
          "data": {
            "text/plain": "Index(['Gender', 'Married', 'Dependents', 'Education', 'Self_Employed',\n       'ApplicantIncome', 'CoapplicantIncome', 'Loan_Amount_Term',\n       'Credit_History', 'Property_Area', 'LoanAmount_log'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test.shape",
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 47,
          "data": {
            "text/plain": "(367, 11)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "X = train.drop('Loan_Status',1) \ny = train.Loan_Status",
      "execution_count": 48,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "X=pd.get_dummies(X) \ntrain=pd.get_dummies(train) \ntest=pd.get_dummies(test)",
      "execution_count": 49,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.columns",
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 50,
          "data": {
            "text/plain": "Index(['ApplicantIncome', 'CoapplicantIncome', 'Loan_Amount_Term',\n       'Credit_History', 'Loan_Status', 'LoanAmount_log', 'Gender_Female',\n       'Gender_Male', 'Married_No', 'Married_Yes', 'Dependents_3',\n       'Dependents_0', 'Dependents_1', 'Dependents_2', 'Education_Graduate',\n       'Education_Not Graduate', 'Self_Employed_No', 'Self_Employed_Yes',\n       'Property_Area_Rural', 'Property_Area_Semiurban',\n       'Property_Area_Urban'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test.columns",
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 52,
          "data": {
            "text/plain": "Index(['ApplicantIncome', 'CoapplicantIncome', 'Loan_Amount_Term',\n       'Credit_History', 'LoanAmount_log', 'Gender_Female', 'Gender_Male',\n       'Married_No', 'Married_Yes', 'Dependents_3', 'Dependents_0',\n       'Dependents_1', 'Dependents_2', 'Education_Graduate',\n       'Education_Not Graduate', 'Self_Employed_No', 'Self_Employed_Yes',\n       'Property_Area_Rural', 'Property_Area_Semiurban',\n       'Property_Area_Urban'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "col_absent_test=[x for x in list(train.columns) if x not in list(test.columns)]",
      "execution_count": 54,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "col_absent_test",
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 55,
          "data": {
            "text/plain": "['Loan_Status']"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Great we have equal columns and same columns in both except the target which is obvious to be absent in test"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from sklearn.model_selection import train_test_split\n\nx_train, x_cv, y_train, y_cv = train_test_split(X,y, test_size =0.3)",
      "execution_count": 56,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## We will use accuracy as metrics"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Model building"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## I) Logistic Regression"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from sklearn.linear_model import LogisticRegression \nfrom sklearn.metrics import accuracy_score\nmodel = LogisticRegression() \nmodel.fit(x_train, y_train)",
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 57,
          "data": {
            "text/plain": "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n          intercept_scaling=1, max_iter=100, multi_class='warn',\n          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n          tol=0.0001, verbose=0, warm_start=False)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "pred_cv = model.predict(x_cv)\naccuracy_score(y_cv,pred_cv)",
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 58,
          "data": {
            "text/plain": "0.8216216216216217"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "pred_test_LOGISTIC = model.predict(test)",
      "execution_count": 61,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Stratified k-fold cross validation\n\nStratification is the process of rearranging the data so as to ensure that each fold is a good representative of the whole."
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from sklearn.model_selection import StratifiedKFold",
      "execution_count": 62,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "i=1 \nkf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True)",
      "execution_count": 63,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "for train_index,test_index in kf.split(X,y):\n    print('\\n{} of kfold {}'.format(i,kf.n_splits))\n    xtr,xvl = X.loc[train_index],X.loc[test_index]\n    ytr,yvl = y[train_index],y[test_index]\n    model = LogisticRegression(random_state=1)\n    model.fit(xtr, ytr)\n    pred_test = model.predict(xvl)\n    score = accuracy_score(yvl,pred_test)\n    print('accuracy_score',score)\n    i+=1\n    pred_test_stratified = model.predict(test)\n    pred=model.predict_proba(xvl)[:,1]",
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "stream",
          "text": "\n6 of kfold 5\naccuracy_score 0.7983870967741935\n\n7 of kfold 5\naccuracy_score 0.8306451612903226\n\n8 of kfold 5\naccuracy_score 0.8114754098360656\n\n9 of kfold 5\naccuracy_score 0.7950819672131147\n\n10 of kfold 5\naccuracy_score 0.8278688524590164\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "##  Feature Engineering\n\n\nBased on the domain knowledge, we can come up with new features that might affect the target variable. We will create the following three new features:\n\n    Total Income - As discussed during bivariate analysis we will combine the Applicant Income and Coapplicant Income. If the total income is high, chances of loan approval might also be high.\n    \n    \n    EMI - EMI is the monthly amount to be paid by the applicant to repay the loan. Idea behind making this variable is that people who have high EMI’s might find it difficult to pay back the loan. We can calculate the EMI by taking the ratio of loan amount with respect to loan amount term.\n    \n    \n    Balance Income - This is the income left after the EMI has been paid. Idea behind creating this variable is that if this value is high, the chances are high that a person will repay the loan and hence increasing the chances of loan approval."
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['Total_Income']=train['ApplicantIncome']+train['CoapplicantIncome'] \ntest['Total_Income']=test['ApplicantIncome']+test['CoapplicantIncome']",
      "execution_count": 68,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "import seaborn as sns\nsns.distplot(train['Total_Income']);",
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "#We can see it is shifted towards left, i.e., the distribution is right skewed. \n#So, let’s take the log transformation to make the distribution normal.\n\ntrain['Total_Income_log'] = np.log(train['Total_Income']) \nsns.distplot(train['Total_Income_log']); \ntest['Total_Income_log'] = np.log(test['Total_Income'])",
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train_original['LoanAmount'].fillna(train_original['LoanAmount'].median(), inplace=True)\ntest_original['LoanAmount'].fillna(test_original['LoanAmount'].median(), inplace=True)",
      "execution_count": 87,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['EMI']=train_original['LoanAmount']/train_original['Loan_Amount_Term'] \ntest['EMI']=test_original['LoanAmount']/test_original['Loan_Amount_Term']",
      "execution_count": 88,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "test_original['LoanAmount'].isnull().sum()",
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 81,
          "data": {
            "text/plain": "0"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['EMI'].isnull().sum()",
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 90,
          "data": {
            "text/plain": "14"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['EMI'].value_counts()",
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 92,
          "data": {
            "text/plain": "0.355556    29\n0.333333    17\n0.305556    16\n0.277778    12\n0.444444    11\n0.519444    10\n0.361111    10\n0.311111     9\n0.416667     9\n0.377778     7\n0.288889     7\n0.263889     7\n0.266667     7\n0.375000     7\n0.347222     6\n0.383333     6\n0.208333     6\n0.319444     6\n0.366667     6\n0.250000     6\n0.500000     6\n0.313889     5\n0.194444     5\n0.388889     5\n0.233333     5\n0.438889     5\n0.222222     5\n0.400000     5\n0.322222     5\n0.555556     5\n            ..\n0.502778     1\n0.702778     1\n3.333333     1\n0.389583     1\n0.761111     1\n0.663889     1\n0.854167     1\n0.700000     1\n0.236111     1\n0.097222     1\n0.969444     1\n0.866667     1\n0.530556     1\n0.297917     1\n0.436111     1\n0.463889     1\n0.441667     1\n0.741667     1\n0.633333     1\n1.266667     1\n0.569444     1\n1.354167     1\n0.822222     1\n0.694444     1\n0.330000     1\n1.388889     1\n0.855556     1\n4.305556     1\n0.172222     1\n0.875000     1\nName: EMI, Length: 237, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['EMI'].fillna(train['EMI'].median(), inplace=True)\ntest['EMI'].fillna(train['EMI'].median(), inplace=True)",
      "execution_count": 93,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['EMI'].isnull().sum()",
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 95,
          "data": {
            "text/plain": "0"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "sns.distplot(train['EMI']);",
      "execution_count": 94,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train['Balance Income']=train['Total_Income']-(train['EMI']*1000) \n# Multiply with 1000 to make the units equal \ntest['Balance Income']=test['Total_Income']-(test['EMI']*1000)",
      "execution_count": 96,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "sns.distplot(train['Balance Income']);",
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 432x288 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Let us now drop the variables which we used to create these new features. Reason for doing this is, the correlation between those old features and these new features will be very high and logistic regression assumes that the variables are not highly correlated. We also wants to remove the noise from the dataset, so removing correlated features will help in reducing the noise too."
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train.columns",
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 99,
          "data": {
            "text/plain": "Index(['ApplicantIncome', 'CoapplicantIncome', 'Loan_Amount_Term',\n       'Credit_History', 'Loan_Status', 'LoanAmount_log', 'Gender_Female',\n       'Gender_Male', 'Married_No', 'Married_Yes', 'Dependents_3',\n       'Dependents_0', 'Dependents_1', 'Dependents_2', 'Education_Graduate',\n       'Education_Not Graduate', 'Self_Employed_No', 'Self_Employed_Yes',\n       'Property_Area_Rural', 'Property_Area_Semiurban', 'Property_Area_Urban',\n       'Total_Income', 'Total_Income_log', 'EMI', 'Balance Income'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "# Now how to get correlation between variables\ntrain[train.columns[1:]].corr()['LoanAmount_log'][:]",
      "execution_count": 100,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 100,
          "data": {
            "text/plain": "CoapplicantIncome          0.206330\nLoan_Amount_Term           0.085353\nCredit_History            -0.018547\nLoan_Status               -0.037997\nLoanAmount_log             1.000000\nGender_Female             -0.143388\nGender_Male                0.143388\nMarried_No                -0.179802\nMarried_Yes                0.179802\nDependents_3               0.123673\nDependents_0              -0.147619\nDependents_1               0.057542\nDependents_2               0.046270\nEducation_Graduate         0.145885\nEducation_Not Graduate    -0.145885\nSelf_Employed_No          -0.109677\nSelf_Employed_Yes          0.109677\nProperty_Area_Rural        0.078614\nProperty_Area_Semiurban    0.007197\nProperty_Area_Urban       -0.083473\nTotal_Income               0.505306\nTotal_Income_log           0.659319\nEMI                        0.435607\nBalance Income             0.483446\nName: LoanAmount_log, dtype: float64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "train=train.drop(['ApplicantIncome', 'CoapplicantIncome',  'Loan_Amount_Term'], axis=1) \ntest=test.drop(['ApplicantIncome', 'CoapplicantIncome',  'Loan_Amount_Term'], axis=1)",
      "execution_count": 108,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## We will try building model again as we have included more features and removed the correlated features\n"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "I will build the following models :\n\n    Logistic Regression\n    Decision Tree\n    Random Forest\n    XGBoost"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "X = train.drop('Loan_Status',1) \ny = train.Loan_Status                # Save target variable in separate dataset",
      "execution_count": 109,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "X.columns",
      "execution_count": 111,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 111,
          "data": {
            "text/plain": "Index(['Credit_History', 'LoanAmount_log', 'Gender_Female', 'Gender_Male',\n       'Married_No', 'Married_Yes', 'Dependents_3', 'Dependents_0',\n       'Dependents_1', 'Dependents_2', 'Education_Graduate',\n       'Education_Not Graduate', 'Self_Employed_No', 'Self_Employed_Yes',\n       'Property_Area_Rural', 'Property_Area_Semiurban', 'Property_Area_Urban',\n       'Total_Income', 'Total_Income_log', 'EMI', 'Balance Income'],\n      dtype='object')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "X.shape",
      "execution_count": 112,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 112,
          "data": {
            "text/plain": "(614, 21)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "X.info()",
      "execution_count": 113,
      "outputs": [
        {
          "output_type": "stream",
          "text": "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 614 entries, 0 to 613\nData columns (total 21 columns):\nCredit_History             614 non-null float64\nLoanAmount_log             614 non-null float64\nGender_Female              614 non-null uint8\nGender_Male                614 non-null uint8\nMarried_No                 614 non-null uint8\nMarried_Yes                614 non-null uint8\nDependents_3               614 non-null uint8\nDependents_0               614 non-null uint8\nDependents_1               614 non-null uint8\nDependents_2               614 non-null uint8\nEducation_Graduate         614 non-null uint8\nEducation_Not Graduate     614 non-null uint8\nSelf_Employed_No           614 non-null uint8\nSelf_Employed_Yes          614 non-null uint8\nProperty_Area_Rural        614 non-null uint8\nProperty_Area_Semiurban    614 non-null uint8\nProperty_Area_Urban        614 non-null uint8\nTotal_Income               614 non-null float64\nTotal_Income_log           614 non-null float64\nEMI                        614 non-null float64\nBalance Income             614 non-null float64\ndtypes: float64(6), uint8(15)\nmemory usage: 37.9 KB\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "X.Property_Area_Semiurban.value_counts()",
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 114,
          "data": {
            "text/plain": "0    381\n1    233\nName: Property_Area_Semiurban, dtype: int64"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Logistic Regression"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "i=1 \nkf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True) \nfor train_index,test_index in kf.split(X,y):     \n    print('\\n{} of kfold {}'.format(i,kf.n_splits))     \n    xtr,xvl = X.loc[train_index],X.loc[test_index]     \n    ytr,yvl = y[train_index],y[test_index]         \n\n    model = LogisticRegression(random_state=1)     \n    model.fit(xtr, ytr)     \n    pred_test = model.predict(xvl)     \n    score = accuracy_score(yvl,pred_test)     \n    print('accuracy_score',score)     \n    i+=1 \n    pred_test = model.predict(test) \n    pred=model.predict_proba(xvl)[:,1]",
      "execution_count": 116,
      "outputs": [
        {
          "output_type": "stream",
          "text": "\n1 of kfold 5\naccuracy_score 0.7983870967741935\n\n2 of kfold 5\naccuracy_score 0.8225806451612904\n\n3 of kfold 5\naccuracy_score 0.7786885245901639\n\n4 of kfold 5\naccuracy_score 0.7868852459016393\n\n5 of kfold 5\naccuracy_score 0.8278688524590164\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "submission=pd.DataFrame()",
      "execution_count": 118,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "submission['Loan_Status']=pred_test            # filling Loan_Status with predictions \nsubmission['Loan_ID']=test_original['Loan_ID'] # filling Loan_ID with test Loan_ID\n\n# replacing 0 and 1 with N and Y \nsubmission['Loan_Status'].replace(0, 'N',inplace=True) \nsubmission['Loan_Status'].replace(1, 'Y',inplace=True)\n\n# Converting submission file to .csv format \npd.DataFrame(submission, columns=['Loan_ID','Loan_Status']).to_csv('Log2.csv')",
      "execution_count": 119,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Decision Tree"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from sklearn import tree\ni=1 \nkf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True) \nfor train_index,test_index in kf.split(X,y):     \n    print('\\n{} of kfold {}'.format(i,kf.n_splits))     \n    xtr,xvl = X.loc[train_index],X.loc[test_index]     \n    ytr,yvl = y[train_index],y[test_index]         \n\n    model = tree.DecisionTreeClassifier(random_state=1)      \n    model.fit(xtr, ytr)     \n    pred_test = model.predict(xvl)     \n    score = accuracy_score(yvl,pred_test)     \n    print('accuracy_score',score)     \n    i+=1 \n    pred_test = model.predict(test) \n    pred=model.predict_proba(xvl)[:,1]",
      "execution_count": 120,
      "outputs": [
        {
          "output_type": "stream",
          "text": "\n1 of kfold 5\naccuracy_score 0.6935483870967742\n\n2 of kfold 5\naccuracy_score 0.7258064516129032\n\n3 of kfold 5\naccuracy_score 0.6967213114754098\n\n4 of kfold 5\naccuracy_score 0.680327868852459\n\n5 of kfold 5\naccuracy_score 0.6967213114754098\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "submission['Loan_Status']=pred_test            # filling Loan_Status with predictions submission['Loan_ID']=test_original['Loan_ID'] # filling Loan_ID with test Loan_ID\n\n# replacing 0 and 1 with N and Y \nsubmission['Loan_Status'].replace(0, 'N',inplace=True) \nsubmission['Loan_Status'].replace(1, 'Y',inplace=True)\n\n# Converting submission file to .csv format \npd.DataFrame(submission, columns=['Loan_ID','Loan_Status']).to_csv('Decision Tree.csv')",
      "execution_count": 121,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "We got an accuracy of 0.63 which is much lesser than the accuracy from logistic regression model. So let’s build another model, i.e. Random Forest, a tree based ensemble algorithm and try to improve our model by improving the accuracy."
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Random Forest"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from sklearn.ensemble import RandomForestClassifier",
      "execution_count": 122,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "i=1 \nkf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True) \nfor train_index,test_index in kf.split(X,y):     \n    print('\\n{} of kfold {}'.format(i,kf.n_splits))     \n    xtr,xvl = X.loc[train_index],X.loc[test_index]     \n    ytr,yvl = y[train_index],y[test_index]         \n\n    model = RandomForestClassifier(random_state=1, max_depth=10)     \n    model.fit(xtr, ytr)     \n    pred_test = model.predict(xvl)     \n    score = accuracy_score(yvl,pred_test)     \n    print('accuracy_score',score)     \n    i+=1 \n    pred_test = model.predict(test) \n    pred=model.predict_proba(xvl)[:,1]",
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "stream",
          "text": "\n1 of kfold 5\naccuracy_score 0.8145161290322581\n\n2 of kfold 5\naccuracy_score 0.7983870967741935\n\n3 of kfold 5\naccuracy_score 0.7213114754098361\n\n4 of kfold 5\naccuracy_score 0.7295081967213115\n\n5 of kfold 5\naccuracy_score 0.8114754098360656\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Lets try using GridSearchCV"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from sklearn.model_selection import GridSearchCV\n\n# Provide range for max_depth from 1 to 20 with an interval of 2 and from 1 to 200 with an interval of 20 for n_estimators \nparamgrid = {'max_depth': list(range(1, 20, 2)), 'n_estimators': list(range(1, 200, 20))}\n\ngrid_search=GridSearchCV(RandomForestClassifier(random_state=1),paramgrid)\n\nfrom sklearn.model_selection import train_test_split \nx_train, x_cv, y_train, y_cv = train_test_split(X,y, test_size =0.3, random_state=1)\n\n# Fit the grid search model \ngrid_search.fit(x_train,y_train)",
      "execution_count": 124,
      "outputs": [
        {
          "output_type": "stream",
          "text": "/home/nbuser/anaconda3_501/lib/python3.6/site-packages/sklearn/model_selection/_search.py:841: DeprecationWarning: The default of the `iid` parameter will change from True to False in version 0.22 and will be removed in 0.24. This will change numeric results when test-set sizes are unequal.\n  DeprecationWarning)\n",
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "execution_count": 124,
          "data": {
            "text/plain": "GridSearchCV(cv='warn', error_score='raise-deprecating',\n       estimator=RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n            max_depth=None, max_features='auto', max_leaf_nodes=None,\n            min_impurity_decrease=0.0, min_impurity_split=None,\n            min_samples_leaf=1, min_samples_split=2,\n            min_weight_fraction_leaf=0.0, n_estimators='warn', n_jobs=None,\n            oob_score=False, random_state=1, verbose=0, warm_start=False),\n       fit_params=None, iid='warn', n_jobs=None,\n       param_grid={'max_depth': [1, 3, 5, 7, 9, 11, 13, 15, 17, 19], 'n_estimators': [1, 21, 41, 61, 81, 101, 121, 141, 161, 181]},\n       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n       scoring=None, verbose=0)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "# Estimating the optimized value \ngrid_search.best_estimator_",
      "execution_count": 125,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 125,
          "data": {
            "text/plain": "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n            max_depth=3, max_features='auto', max_leaf_nodes=None,\n            min_impurity_decrease=0.0, min_impurity_split=None,\n            min_samples_leaf=1, min_samples_split=2,\n            min_weight_fraction_leaf=0.0, n_estimators=141, n_jobs=None,\n            oob_score=False, random_state=1, verbose=0, warm_start=False)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "i=1 \nkf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True) \nfor train_index,test_index in kf.split(X,y):     \n    print('\\n{} of kfold {}'.format(i,kf.n_splits))     \n    xtr,xvl = X.loc[train_index],X.loc[test_index]     \n    ytr,yvl = y[train_index],y[test_index]         \n    model = RandomForestClassifier(random_state=1, max_depth=3, n_estimators=41)     \n    model.fit(xtr, ytr)     \n    pred_test = model.predict(xvl)     \n    score = accuracy_score(yvl,pred_test)     \n    print('accuracy_score',score)     \n    i+=1 \n    pred_test = model.predict(test) \n    pred2=model.predict_proba(test)[:,1]",
      "execution_count": 126,
      "outputs": [
        {
          "output_type": "stream",
          "text": "\n1 of kfold 5\naccuracy_score 0.7983870967741935\n\n2 of kfold 5\naccuracy_score 0.8225806451612904\n\n3 of kfold 5\naccuracy_score 0.8032786885245902\n\n4 of kfold 5\naccuracy_score 0.7622950819672131\n\n5 of kfold 5\naccuracy_score 0.819672131147541\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "submission['Loan_Status']=pred_test            # filling Loan_Status with predictions submission['Loan_ID']=test_original['Loan_ID'] # filling Loan_ID with test Loan_ID\n\n# replacing 0 and 1 with N and Y \nsubmission['Loan_Status'].replace(0, 'N',inplace=True) \nsubmission['Loan_Status'].replace(1, 'Y',inplace=True)\n\n# Converting submission file to .csv format \npd.DataFrame(submission, columns=['Loan_ID','Loan_Status']).to_csv('Random Forest.csv')",
      "execution_count": 128,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "##  Feature importance"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "importances=pd.Series(model.feature_importances_, index=X.columns) \nimportances.plot(kind='barh', figsize=(12,8))",
      "execution_count": 129,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 129,
          "data": {
            "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x7fd0adaa0a90>"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": "<Figure size 864x576 with 1 Axes>"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "We can see that Credit_History is the most important feature followed by Balance Income, Total Income, EMI. So, feature engineering helped us in predicting our target variable."
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## XGBOOST"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from xgboost import XGBClassifier\n\ni=1 \nkf = StratifiedKFold(n_splits=5,random_state=1,shuffle=True) \nfor train_index,test_index in kf.split(X,y):     \n    print('\\n{} of kfold {}'.format(i,kf.n_splits))     \n    xtr,xvl = X.loc[train_index],X.loc[test_index]     \n    ytr,yvl = y[train_index],y[test_index]         \n    model = XGBClassifier(n_estimators=50, max_depth=4)     \n    model.fit(xtr, ytr)     \n    pred_test = model.predict(xvl)     \n    score = accuracy_score(yvl,pred_test)     \n    print('accuracy_score',score)     \n    i+=1 \n    pred_test = model.predict(test) \n    pred3=model.predict_proba(test)[:,1]",
      "execution_count": 132,
      "outputs": [
        {
          "output_type": "stream",
          "text": "\n1 of kfold 5\naccuracy_score 0.7903225806451613\n\n2 of kfold 5\naccuracy_score 0.8225806451612904\n\n3 of kfold 5\naccuracy_score 0.7622950819672131\n\n4 of kfold 5\naccuracy_score 0.7704918032786885\n\n5 of kfold 5\naccuracy_score 0.7950819672131147\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "submission['Loan_Status']=pred_test \nsubmission['Loan_ID']=test_original['Loan_ID']\n\nsubmission['Loan_Status'].replace(0, 'N',inplace=True) \nsubmission['Loan_Status'].replace(1, 'Y',inplace=True)\n\npd.DataFrame(submission, columns=['Loan_ID','Loan_Status']).to_csv('XGBoost.csv')",
      "execution_count": 133,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### More techniques to be tried are:\n\nGridSearch on XGB\nDemo on different hyper parameter opt on XGB\n\nI) Hyperopt\n\nII) Spearmint\n\nIII) SMAC\n\nTry more algorithms like\n\nI) Catboost\n\nII) LightGBM"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "name": "python36",
      "display_name": "Python 3.6",
      "language": "python"
    },
    "language_info": {
      "mimetype": "text/x-python",
      "nbconvert_exporter": "python",
      "name": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.6",
      "file_extension": ".py",
      "codemirror_mode": {
        "version": 3,
        "name": "ipython"
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}